Description: (from applicant's Abstract) Recent exploration into the role of the immune system, genetic susceptibility, and environmental factors such as infections, vaccines, and diet has led to the emergence of numerous theoretical models to explain the biological basis for autism. In turn, a growing number of parents are implementing a variety of still unproven treatment modalities with their autistic children. Particularly relevant to our proposed research is recent work by Cade and colleagues on diet and autism. Testimonials throughout the world support the efficacy of a diet free of wheat (which contains gluten) and milk (which has casein). Most noteworthy are reports of autistic children who, after being placed on the restricted diet, are "cured" of their autism. However, there are no published reports of controlled clinical trials using the most widely accepted autism diagnostic instruments and direct behavioral observation to test dietary effects. In this proposed research, we will use a randomized, double-blind repeated measures crossover design to evaluate the recommended gluten-free and casein-free (GCF) diet in a controlled inpatient setting. The specific aims are: (a) to evaluate the effects of the GCF diet on severity of autistic symptoms as measured by the Childhood Autism Rating Scale (CARS), Ecological Communication Orientation Scale (ECOS), and direct behavioral observation frequencies, (b) to evaluate the effects of the GCF diet on urinary peptide levels and (c) to evaluate the role of parent behavior in therapeutic and placebo effects of the GCF diet in autistic children. We hypothesize that: (a) using the placebo dietary condition for comparison, scores on the CARS, ECOS, and direct in-home observation ratings (frequency counts) of child social and language behaviors and aberrant behaviors will show an improvement following completion of 5 weeks on the GCF diet, (b) adherence to the GCF diet will result in decreased urinary peptide levels, and (c) "expectant waiting" behavioral frequencies, a measure of positive expectancy and subjectivity on the part of the parent, will not increase at any point during the study. Results of this study should provide important preliminary, empirically sound data regarding current claims of positive social and language behaviors in autistic children.